The long-term objective is to investigate determinants of breathing instability in patients with sleep apnea. The central hypothesis is that patients with sleep apnea demonstrate ventilatory control instability, that is either a cause or a consequence of sleep apnea. In addition, ventilatory control during sleep is not immutable, but demonstrates substantial plasticity. Long-term facilitation following episodic hypoxia is a known example of plasticity that may increase chemoresponsiveness and potentially aggravate breathing instability during sleep. These hypotheses will be tested in patients with obstructive sleep apnea and normal controls. The following specific aims will be addressed: Specific Aim 1 is to determine the hypocapnic apneic threshold in patients with obstructive sleep apnea. We will use non-invasive nasal mechanical ventilation during sleep to test the hypotheses that the hypocapnic apneic threshold is elevated in patients with obstructive sleep apnea. Specific Aim 2 is to determine whether ventilation is under behavioral rather than chemoreflex control during wakefulness in patients with OSA. We will use rebreathing during wakefulness under hypoxic and hyperoxic conditions to test the hypothesis that the chemoreflex threshold is near or above eupneic PE-CO2 levels and chemoreflex sensitivity is greater in subjects with moderate sleep apnea compared to matched controls. Specific Aim 3 is to determine the effect of long- term facilitation (LTF) on ventilatory stability during sleep. We will induce central apnea with nasal mechanical ventilation before and after episodic hypoxia to test the hypothesis that LTF following repetitive hypoxia elevates the hypocapnic apneic threshold during NREM sleep and hence may increase susceptibility to develop sleep apnea and breathing instability during sleep. The proposed protocols are likely to yield critical insight into the pathogenesis of breathing instability during sleep and may point toward effective pharmacologic therapy for sleep apnea.